Subject: Re: silo and packets: overflowing too-bigness
To: Martin Husemann <martin@rumolt.teuto.de>
From: Bill Sommerfeld <sommerfeld@orchard.arlington.ma.us>
List: tech-kern
Date: 07/31/1998 11:13:55
> Should these options be stripped by the kernel when the packet is queued
> for a point-to-point interface? This way only your router needs to take
> care of this, now you have to disable it on all machines on your local
> network.
Simply stripping options won't work, because (for instance) the window
scale option *changes the meaning* of some of the other fields in the
header..
It also would require in-flight modifications to the tcp checksum,
causing it to no longer be an end-to-end checksum..
- Bill